Improving few-shot object detection by saving and hallucinating examples
Weilin Zhang · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · 2020
Learning to detect an object in an image from very few training examples - few-shot object detection - is challenging, because the classifier that sees proposal boxes has very little training data. A particularly challenging training regime occurs when there are one or two training examples. In this case, if the region proposal network (RPN) misses even one high intersection-over-union (IOU) training box, the classifier's model of how object appearance varies can be severely impacted. We use multiple distinct yet cooperating RPN's. Our RPN's are trained to be different, but not too different; doing so yields significant performance improvements over state of the art for COCO and PASCAL VOC in the very few-shot setting. This effect appears to be independent of the choice of classifier or dataset. However, under the very low-shot regime, even if all high IOU boxes are used to train the classifier, the variations are still insufficient to train the classifier in novel classes. We propose to build an even better model of variation in novel classes by transferring the shared within-class variation from base classes. We introduce a hallucinator network and insert it into a modern object detector model, which learns to generate additional training examples in the Region of Interest (ROI's) feature space. This approach yields further performance improvements on two state-of-the-art few-shot detectors with different proposal generation processes.